Fired OpenAI Researchers Say Safety Concerns Led to Exit — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Mikita Balesni, Tomek Korbak, and Jasmine Wang were fired from OpenAI and claim their dismissals resulted from prioritizing AI safety over corporate interests, not from mishandling data.
- OpenAI stated its internal investigation confirmed the researchers mishandled sensitive information outside company procedures and called the decision to terminate them a matter of breached trust.
- OpenAI research leaders denied the firings were related to safety concerns, asserting in a public note that the decision was based on a 'significant breach of trust' beyond what the employees disclosed.
- Tomek Korbak had raised alarms internally about losing the ability to monitor AI agents' internal reasoning, which he believes led to his dismissal.
- Jasmine Wang warned on X that the firings set a dangerous precedent, citing past suspicious departures and urging employees to resist what she sees as a pattern of silencing safety voices.
- OpenAI is finalizing contracts with third-party safety assessors, with details expected in the coming weeks, as part of its response to growing scrutiny over internal safety practices.
Why it matters: Current and future AI safety researchers may now hesitate to raise concerns, risking a culture of silence within a company shaping powerful technologies. OpenAI’s claim of a 'breach of trust' without public evidence creates tension between its safety mission and corporate control, with material consequences for employee morale and public trust.
Ask SkimNews
